Okvir za interoperabilnost podatkov senzorjev v pametnih okoljih

This final project paper examines whether semantic interoperability for heterogeneousIoT sensor data can be achieved without relying on centralized cloud infrastructure or heavyweight ontology reasoning.To investigate this, a lightweight edge-mediated framework based on a profile-based architecture was designed and implemented.The system performs syntactic normalization, structural alignment, and semantic annotation locally using JSON-LD with standard vocabularies (QUDT, W3C IoT) at the edge.Because mediation logic executes on local edge hardware, the approach supports resource-efficient and cloud-independent data integration.The system was evaluated in experimental performance tests across five test runs with 100 sensor messages each.The results showed that the framework achieved high throughput (585.11msg/sec), efficient memory usage (79.1 MB), acceptable latency (89.40 ms median), and perfect semantic correctness (100%).Mitrinovska M. A Framework for Sensor Data Interoperability in Smart Environments.